Will Lithium Really Save Your Brain?

There is an element sitting in your brain right now, and nobody has established how much you are supposed to have.  It may shape how you think, how you act, and whether your mind holds up as you age. Last year a Harvard lab placed it near the center of Alzheimer's, arguing that running short of it may be one of the first things that goes wrong in the diease. It is already sold as the next longevity compound, the thing that finally protects the aging brain, and the people selling it have a real paper behind them. This investigation takes that case apart study by study to find out how much of it holds. The question sounds simple. The answer, as always, is not.

00:00 – Intro: Lithium and the Brain
01:57 – Lithium's Origin as a Bipolar Drug
04:55 – The Brain Growth Claim
10:10 – Where the Evidence Breaks Down
12:12 – The Harvard Brain Study
15:36 – Should You Take Lithium?
18:04 – New Trial Starting This Fall

Transcript

Copy Transcript Copied

SPEAKER_00 0:00

There is an element sitting in your brain right now. Nobody has ever established what a normal mountain looks like, and there is no reference range for it, but it may be shaping how you think, how you act, and whether your mind holds up as you age. Last August, a laboratory at Harvard Medical School published a paper, Nature, arguing that this element sits close to the center of Alzheimer's disease, and that running short of it may be one of the first things that goes wrong. A randomized trial at the University of Pittsburgh published this year gave that same element to people with early memory loss for two years, again, six outcomes it had named in advance. None of them cleared the bar it had set. Both are serious work, and this month a psychiatric trade publication wrote that failed trial up as a success, which tells you something about how badly people want one of these answers over the other. Millions are already taking it off the shelf as a supplement. It's being sold to something close to an anti-aging compound, the thing that finally protects the aging brain. Questions whether that holds, whether this is genuinely the next level anti-aging compound that saves your brain, or something narrower that got promoted well past what the evidence supports. I am William Wallace, and this is the Active Site. In March of 1949, an American medical journalist published two reports back-to-back on the same issue. Heart patients had been told to cut sodium, so manufacturers sold them a salt substitute, and nobody knew how much anyone was swallowing because there was no way to measure it. They salted their food freely, people were poisoned, and some of them died. Two weeks before those papers appeared, the Food and Drug Administration had already moved against the products. Six months later, in September, an Australian psychiatrist published something else. He had given the same compound to men who spent years locked away with mania, and several of them recovered well enough to go home. Three-quarters of a century later, and it's still a first-line treatment for bipolar disorder, which is something very rare for a natural compound. Within a single year, the same element was written up as a poison and as a cure. It was lithium. The psychiatrist was not introducing a foreign chemical into those men. Lithium is in grains and vegetables and tap water, and it had been detected in human tissue since the 1920s. It was already in every one of those patients before he treated a single one of them, and in him, and it's in you right now. The only question was if nobody could answer that in a clinic because no test existed that was built to. Lithium had been found in human blood as early as 1931, but nothing in that literature was built to say what a normal amount is. What arrived in 1950 was a way to measure lithium, and somebody who had taken too much developed as a direct response to the poisonings. Most of what a person takes in comes from food, from grains and vegetables. Some of it comes from water, and that is the part that varies because lithium leaches out of rock into groundwater, and how much reaches a tap depends entirely on what the water ran through on the way. Across a single country, the levels vary more than fifty fold, which turns the map into something close to an experiment nobody designed. Geography had sorted out whole populations into different water concentrations decades before anyone thought to check the results. In 1990, a researcher named Gerhard Schrouser lined up county after county in Texas against the lithium in their water supply and found that places with the least of it had significantly higher rates of violent crime and self-harm. It was easy to dismiss as a fluke of one state, but then it kept happening. In Japan in 2009, across Austria in 2011, back in Texas in 2013, and in Greece just a few years later. Four research teams, four countries, and more than two decades. That is an eerily consistent finding in epidemiology. Now, behavior is one thing, but that leads to an obvious question. What is a lifetime of lithium exposure doing to the brain itself? In 2017, a team at the University of Copenhagen in Denmark went looking, they modeled the lithium in drinking water across the entire country, then they took every Dane ever diagnosed with dementia and matched each one against 10 people who had stayed healthy. Among those drinking the most lithium, dementia was 17% less common. The medical records point the same way. A study in 2024 found that people prescribed lithium for psychiatric reasons had an Alzheimer's risk around 40% lower than people who never were. A study in 2026 found closer to half, all of which raises a question about what this element actually is. Is it a drug or is it something closer to a nutrient you're supposed to have a certain amount of? The nutrition literature leaning toward possibly an essential nutrient. Rats and goats raised on low lithium feed have higher mortality and reproductive problems, which is the standard pattern for something a body requires. A review in 2002, published by the same chemist, Gerhard Schrauser, the same man who published the first evidence of lithium in Texas drinking water in 1990, laid out that case and proposed a daily requirement of about one milligram based on that animal mortality data, which would make it a nutrient with a requirement, like iron or zinc, and a number you could be under if you weren't getting enough. And then there's a seemingly strong anti-aging case for it. Start with the shape of the brain itself. Bipolar disorder shrinks the hippocampus, the seahorse-shaped structure that files new memories. In 2012, a team pulled imaging studies and split the patients by whether they were taking lithium. The ones who were not had hippocampi measurably smaller than healthy people. The ones who were had hippocampi larger than healthy people. Same illness, same scanners, and a very large gap between the two. There was more. In 2000, a study in the Lancet scanned bipolar patients before and after a month of lithium and reported that the volume of their gray matter had increased. A randomized study a decade later found the same thing, and the patients whose gray matter grew the most were the ones who responded best to lithium treatment. The obvious reading was that lithium builds brain tissue, then the part that turned it into a longevity story. In 2011, a nutrition group at the University of Jena in Germany compared tap water lithium against death from any cause across a million people in Japan and found that where there was more lithium, fewer people died of anything. So by this point, the case looks close to complete, an element already inside you that preserves the structure of the brain that appears to grow actual brain tissue and tracks with less dementia across whole countries, that tracks with people living longer, and that animal work says a body may actually require. If that holds, it's the closest thing to a genuine anti-aging compound anyone has found. It costs almost nothing and it has been sitting in our water supply the whole time. That is the argument, anyway, and it's the strongest version of it. Every study we mentioned to this point is real. But now watch what happens when each piece of it's tested on its own. Start with the most direct test there is. Give lithium to people, give a placebo to people who look just like them and see who does better. In 2009, six academic memory clinics gave lithium at a full psychiatric dose to patients with mild Alzheimer's disease for 10 weeks. Neither of the things it set out to measure moved. The University of Pittsburgh trial that was just published tried the opposite. A low dose well under the psychiatric range given for two full years against six outcomes it had named in advance. On memory, the group taking lithium declined at half the rate of the group taking nothing, but before the trial began, the researchers had set a bar the result needed to clear to count, and it did not clear it. Of the other five measures, four showed nothing at all, and the fifth pointed the same way as in memory without getting there either. A full dose for ten weeks in the 2009 study and nothing, a much smaller dose given for two years in the Pittsburgh study, and a result that was objectively disappointing. Just a few days ago, a trade publication took the results of that Pittsburgh trial and wrote it up as a success. Not by inventing anything, by taking the one number that leaned in the right direction, reporting it without the threshold it failed to clear, and letting the headline do the rest. These were the studies that should have settled whether lithium was working in the brain of Alzheimer's people, and they could not. And they were not the only thing that thinned out. Every pillar of the longevity case gives way the same manner once somebody tests it on its own. Take the gray matter in the brain. In 2013, a team at Newcastle University tested whether lithium actually grows tissue or only appears to. They gave healthy men lithium or a placebo that measured brain volume two independent ways. One method showed the increase, the other showed nothing at all. Lithium also changed how gray matter responds to a magnetic field, which means it changed how the tissue photographs. The authors were careful and said only that the difference might come from the signal rather than from actually building the brain. Nobody has ruled that out since. The clock slowing claim, the longevity one, it thinned the same way. In 2024, researchers at King's College London went into the UK Biobank, found the people who actually had been prescribed lithium, and tested how long they had taken it against the length of their telomeres, the protective caps on chromosomes that shorten as a person ages. Three separate measures of exposure, no association with any of them. Then the essential nutrient argument. Go back to that 2002 review by Gerhard Schrauser, the source of the recommendation for one milligram a day, and read into it further. The same paper states plainly that no lithium deficiency disease has ever been characterized in a human being. The animals are real, the requirement is an extrapolation from them, and no expert panel has ever adopted it. Now let's look at all the water studies we mentioned. The pattern did not hold everywhere. A study across the east of England found nothing, one in Portugal found nothing, one in Argentina where the water carries far more lithium pointed the other way entirely and suggested harm. When somebody pulled the whole literature, there was a telltale gap in it. Small studies that found a big effect were there, small studies that found nothing were missing. That pattern usually means those studies were run and never published, because a result showing nothing is much harder to get to a journal. Then the largest test of all. Three and a half million Danes tracked as individuals rather than as regions followed for 22 years. The same research group using the same water map that produced the dementia finding in Denmark. And across every level of lithium that Danish water actually contains, they found nothing. That is the pattern, and once you see it, it's everywhere in this story. At the level of the population, the signal is large and consistent, follow individual people with known exposure, and it fades to nothing. So where does that leave us to this point? The brain tissue growth may be a trick of the scanner. The clock slowing in anti-aging shows nothing in people who actually took it. The longer lives are a comparison between places, not people. Lithium might not be a proven anti-aging compound, and the two mechanisms that were supposed to make it one are the two that failed the hardest. But something did survive despite all of that. Go back to the trials because they share something, and it's not a flaw in how they were run. It's something that they all assumed before they even began. Each of them took people, gave half of them lithium, and weighted. Almost none first measured how much lithium those people already had. One did. A small trial in Dundee, and the result is the most useful number in this entire story. Every single baseline reading came back below the detection limit of the test. The instrument could not resolve the thing the trial needed to know, and no trial has ever used a person's own lithium level to decide whether to give them any. That sounds trivial, but a baseline is the most ordinary measurement in medicine. The number you take out of reflex, hand out reading glasses to everyone in a room and then measure how much the reading improves, and the answer comes back close to zero. Not because reading glasses don't work, because most of the room could already see, and a handful of people who needed them get averaged into everyone who did not. Run that experiment a dozen times, and you will conclude that reading glasses do nothing. Which is a strange thing for an entire field to overlook until you look at what they had to take that measurement with. The blood test measures blood and it was built to find a ceiling. The brain was never on the table at all. For decades, researchers have taken Alzheimer's brains apart, screening them for every metal anyone thought to name. Lithium was on none of those panels. It was never a suspect, so it was never ruled out. Lithium can be seen inside a living skull, but the machines used to scan for metals in the brain were historically never sensitive enough to catch the trace amount a brain carries on its own in lower concentrations, and every scan of that kind was done on someone who had been given lithium first. No one had ever looked at an ordinary brain and asked how much lithium naturally belongs there. In 2025, somebody finally did. Harvard Medical School took frozen tissue from the prefrontal cortex of people who had died with their memory intact, people who had died with Alzheimer's, and a group in between whose memories had only just begun to slip. 27 metals in every sample, a wider panel than anything that had come before it. One metal was lower in both the middle group and the diseased group. It held in the second set of brains, it held when another laboratory ran the samples, it held under a different preparation method, and a control ruled out the obvious artifact. That metal was lithium. Then they asked where it had gone, lifting tissue away in tiny patches with a laser and sorting the patches by whether they held amyloid plaque, the protein deposit that builds up in Alzheimer's. Some of the lithium was inside the plaque, several times more concentrated than the tissue beside it. And of the lithium still outside the plaque in the fraction a cell could actually reach, there was less again. So there is less lithium in an Alzheimer's brain to begin with, and a larger share of what remains is stuck to something that will not give it back. It's like a bank account that is shrinking and partly frozen at the same time, which raises an obvious objection. These amounts are tiny, far below anything a doctor would prescribe. Why would running short of something that small do any damage at all? Picture a rail line running the length of a nerve cell, because that is close to what this is. A neuron is long and thin, and everything it needs has to travel from one end to the other. Supplies move along internal tracks, and those tracks are held in place by a protein called tau, which works like the wooden ties fixing the rails to the ground. There is an enzyme in every neuron whose job it is to loosen those ties. That sounds destructive, but it isn't. A rail line needs maintenance, sections need lifting and drilling, and the enzyme is the crew that does it. In a healthy brain, the crew works at a measured pace and the line stays open. Lithium is one of the things setting that pace. Take it away and the crew does not stop, it works harder. The ties come up faster than anything can put them back, and once enough are gone, towelets go with a rail entirely. Loose tau curls in on itself and clumps together with other loose tau, and those clumps are the tangles found in the brains of people who died with Alzheimer's. The line comes apart, delivery stop, and the far end of the cell, the end furthest from the supply, starves first. The enzyme is called GSK3 beta. In a lithium-depleted brain, there is not just more of it working, there's more of it. Depletion raised both the amount in the enzyme and the form that switches it on. How lithium holds it back is still unsettled, and the authors say so. They put healthy mice with no genetic risk on a lithium-strip diet, cutting brain lithium by about half. Those animals developed elevated soluble amyloid inflamed immune cells, loss of dendritic spines, loss of axons, and failing memory. What they did not develop was plaques or tangles, those required mice already carrying Alzheimer's genes. And the reversal was a separate experiment. They picked the salt that holds together best, and water, on the theory, would be captured by the plaque less. That salt is lithium oritate, the same compound sold in capsules as supplements. It went into the drinking water of mice bred to develop Alzheimer's and of aging normal mice, none of which had been on the strip diet. It cut their plaque burden sharply, but nobody in the paper took a lithium-starved animal and gave the lithium back. What did reverse the damage from depletion was a drug that blocks GSK3 beta directly with no lithium involved at all. And every human brain in this story was measured once after death. Nothing about what came first can be established from it, and the paper's own wording stays careful. Lithium loss may be an early event, but maybe not the first one. So is lithium the next level anti-aging compound some have built it up to be? Well, not on this evidence, but that's a different question from whether it does anything, whether it's actually useful, and the two have different answers. The people taking it are not chasing nothing. They're chasing hippocampi that hold their size, water supplies a track with longer lives, and a metal that goes missing from the brain before an Alzheimer's diagnosis arrives. Anyone dismissing this as a psychiatric drug with a bad reputation is arguing against a very large body of evidence. What that evidence cannot do yet is follow one person. Nearly all of it compares places to places, or groups to groups, and every time somebody tracked individuals instead, the effect shrank or vanished. The gray matter, the telomeres, the largest water study ever run, those were the pillars the anti-aging claim stood on, and those are the ones that fell. What did not fall is the dementia signal. Hundreds of thousands of people who took lithium for other reasons got less dementia, and the measurement showing why has been replicated three times over. On that specific question, is lithium useful? May it possibly have a role in dementia prevention? I would say maybe yes, and I would not say that about most things I look at. What is not supported is a dose. To correct a deficiency, somebody first has to have a defined one and nobody has. No requirement, no range, no test. The one figure in circulation traces back to a paper that says in its own text that no human deficiency has ever been characterized. Anyone selling you a dose is selling you a number that does not exist yet. The form is not settled either. Lithium orotate has been sold since the 1970s on the idea that the orotate carries minerals into cells intact, a claim shown for other minerals and assumed for this one. Nobody has ever found an intact lithium orotate molecule in a living body, and a paper last month argues the bond comes apart in the stomach acid long before it could, which leaves where I land personally. Brains on their way to Alzheimer's carry less usable lithium than healthy brains, and it's already true before anyone gets a diagnosis. The field spent three-quarters of a century unable to see that because every instrument it owned was pointed somewhere else. What runs short is not a supplement, it's a thing holding back a crew that will otherwise tear up track faster than a cell can lay it back down. If you are taking lithium, you are betting on a mechanism that looks right rather than a result anyone has demonstrated, and you're picking your own dose because nobody has established a correct one. That is a defensible bet, but it's not a proven one. The missing study starts this October. Johns Hopkins is due to begin the first randomized placebo controlled trial of lithium orotate in Alzheimer's disease with spinal fluid sample before and after to find out whether a swallowed dose reaches the nervous system at all. For three-quarters of a century, the question was whether lithium works. Nobody was ever able to ask who needs it. Somebody is finally asking. Thank you, everyone, who stayed to the end. If you learned something, please like and subscribe wherever you're listening or watching this. We publish investigations every Tuesday. Please drop a comment and let me know what you would like investigating next. Until next time. Stay healthy.